Inside Track: The Greatest Showman ‘Rewrite The Stars’

Also heavily involved in the musical side of the production was Alex Lacamoire (left), with Greg Wells in this photo from Rocket Carousel.

Greg Wells has worked with everyone who’s anyone in the world of music — but nothing could prepare him for mixing a film soundtrack!

Greg Wells has nearly 30 years of top-level experience as a musician, songwriter, engineer, mixer and producer. He works in a wide range of genres, and plays drums, keyboards, guitar, and several other instruments, with A-list credits including Elton John, Katy Perry, Adele, Timbaland, Pharrell Williams, Celine Dion, Pink, Dua Lipa and many more. However, until he began work on the Original Motion Picture Soundtrack for The Greatest Showman, Greg Wells had never worked on a movie before. The experience would force him to completely change a production approach developed over decades, and left his cherished collection of vintage outboard gathering dust.

“In the middle of 2016,” says Wells, talking from his Rocket Carousel Studio in Los Angeles, “I had a meeting with film director Michael Gracey, here in my control room. I believe I got on his radar because he was a fan of my work, and he felt that what I did musically corresponded with the visual aesthetic he was pursuing with the movie, which was to make a modern contemporary movie about events that happened in the mid-1800s. It had to look like a period piece, yet they also wanted it to be fresh. He acknowledged it’s potentially a dangerous concoction, because it’s easy to be goofy and ridiculous when you have modern choreography and pop songs in a movie set in 1850.

“According to Michael, the things I do as a record producer pay homage to the classics, but they also are fresh and contemporary, without being gimmicky. This is indeed important to me. I’ve never gone for gimmicky sounds that are going to be dated in two years from now. Following the fickle trends of fashion, whether in clothing, music, or whatever, has never really interested me. Things date well when you try not to follow that stuff. Instead I just try and present music in the most vital, true way, and Michael liked that about my approach. While he was here, he stood in front of his small laptop and acted out the entire movie for me in two hours, orating the lines of the different characters. It was an amazing, crazy meeting, and I was completely sold. I loved it, and him!”

Insanity Beckons

Greg Wells in his own Rocket Carousel Studio, where most of his work on The Greatest Showman took place.For almost a year, however, Wells heard nothing more. Eventually, after a brief hold-up to finish another project, he received the Pro Tools session for a song called ‘From Now On’, which he produced and mixed using his established way of working, incorporating the enormous amount of classic analogue gear in his studio and his Eric Valentine-designed 24-channel valve Undertone Audio desk. It was at this point that things began to get more complicated...

“20th Century Fox had hired several very talented music producers, who were all at the top of their game, but the movie producers found that as much as they loved what these people had done, they did not end up with the choices and changes that were required. From my end, I have a wide range of experience as a musician, playing in symphony orchestras, punk bands, weddings, and so on. I recall once, when I was 19, driving 10 hours north of Toronto to perform with a Dixieland band playing at the opening of a take-out French Fries counter. I was dressed in a goofy tuxedo and had a guy dancing next to me in a French Fry costume. Naturally, people my age were walking by shaking their heads. But I loved it!

Team photo: from left, composers Benj Pasek and Justin Paul, actor Hugh Jackman, Greg Wells, and director Michael Gracey.“So I have all these diverse experiences that I bring to situations like this, and I indeed needed every bit of experience that I have, because they had been working on the songs for this movie for four years, and had kept everything that had been done. As a result the average track count of the sessions was 350-400 tracks! When they gave me the first song, the director told me that I had full control and that I could use everything that was in the session, or some of it, or nothing and start again — anything, as long as I kept the vocals.

“Every session was very organised, and the top tracks would be from demos recorded years and years ago, done by one producer, and then there would be orchestral session number one done by another producer, then orchestral session number two by yet another producer, a horn section, another horn section, perhaps a third horn section, plus different choirs from different sessions recorded in different cities, and so on!

“It was absolute insanity, and what was required was for me to make big choices and sweeping changes, rather than just say: ‘That snare sound is not bright enough.’ Instead, I ended up doing huge things like rewriting string parts, deleting entire sections, replaying a bass guitar because the chords in the song had been changed since the original bass part had been recorded, adding guitars, piano, lots of live and programmed drums, and so on. I brought all those skills to the table, because for me it’s all part of the same brush stroke. Thankfully they liked what I did on that first song, because they then gave me another two songs to work on, and then four, and then six, and eventually I produced and mixed most of the album.”

More & More

A month was spent in Sweden, working at Riksmixningsverket, the Stockholm studio of Abba super-producer Benny Andersson. Justin Paul (left) and Benj Pasek debate a point of arrangement.Wells completed his analogue mix of ‘From Now On’ in June 2017, and then set to work on the other tracks in a process that lasted half a year. This took place entirely in his own studio, apart from a month spent in Stockholm (because his wife is Swedish), at Benny Andersson’s Riksmixningsverket Studio. When Wells embarked on work on the second song, it became clear that he could not continue the way he had done, and that his entire approach needed to be overhauled. The almost unmanageable amount of tracks was one reason, the vociferous needs of what he calls “the movie beast” another.

“Actually, a few songs had much higher track counts than 400! I was very intimidated by this for about a week. But once you start to understand something, it becomes smaller and easier to digest. But the over-the-top, macro nature of it all remained really daunting to me. I remember saying to the two songwriters for the movie, Justin Paul and Benj Pasek, ‘Normally I find that less is more,’ and they laughed and said, ‘Greg, we think that in this project, you’ll find that more is actually more.’ So that kind of became the motto of the whole project.

The console at Riksmixningsverket, with Benny Andersson’s custom white Minimoog.“There is so much going on in these sessions, and I had to wrap my head around the best way to keep everything organised, where I could quickly turn all the drums off from, say, [producer] Ricky Reed’s element of a production, or turn them back on, or solo something. Often when working on these songs I could hear a track, but just couldn’t find it in the session, because there were hundreds and hundreds of tracks. It was so confusing. Sometimes that happened every five minutes, and everyone would be laughing: ‘Greg can’t find the track again!’ But hey, they couldn’t either! Nobody could! Plus some producers had used incredibly complex, insane routing, particularly on some of the orchestral elements, and the only solution was to leave those routings, and build my own mix template around them.”

High Tension

The process of getting the songs into shape was, of necessity, a collaborative one. “Alex [Lacamoire] was very involved during the Summer, and then he had to go back to Hamilton-land and deal with the opening of the musical in London. The person who I worked most closely with throughout was Justin [Paul] who is more the musician in the writing duo, while Benj is more on the lyric side. There was no regular way in which we approached things. We all just showed up in the morning and said to each other: ‘What shall we work on today?’ Sometimes we worked on several songs on one day. Justin drove the project in terms of what we focused on, not only because he had worked on this project for four years, but also because he has a fine-tuned brain that allows him to remember everything they had recorded. He has folders inside of folders inside other folders on his laptop, and he’d remember every single element of every single version, even if it was done three years ago.

“A lot of the time it was Justin and me hunkered down, glued to the screen, trying to create something neither of us could on our own. Plus, he plays like a high voltage cable is running through him, and I wanted the music to sound like the way he looks when playing. I felt that the musicians on some of the recordings did not always deliver the kind of intensity that is naturally coming through him, and so I got Justin to overdub guitar and keyboards. In terms of how I approached the music production, I did not think of the 1850 setting of the movie. But I did always work to picture. I insisted that the movie was always playing while we worked on the music. I imagined that I was in the movie theatre and I was trying to make the music match the visuals. The visuals are stunning and the performances of the actors are incredible and full of vitality. Director Michael Gracey is a genius with visuals. The music that was sent to me did not quite match that, and much of what I did was to try to make things feel energetically right coming out of the speakers.

“Speakers are brutal, and we had to jump through all these hoops and do all these crazy techniques to make the music zing with electricity and make it sound like a great live performance. Looking at a live performance of a beautifully shot, beautifully edited, beautifully directed big-budget movie dictated the height of the bar. It took a while to match that, but we got there, even though it was a Herculean project that kind of drove me a bit crazy by the end. The stress level was much higher than I am used to, though it got to the point where that was kind of normal, and afterwards it took me about a month to come down from the intensity.”

The Greatest Showman dramatises the life of legendary impresario PT Barnum, played by Hugh Jackman.

Star Turn

To illustrate the hoops he jumped through and the “crazy techniques” he developed to get the music to the required standard, Greg Wells focuses on one of the singles from The Greatest Showman, ‘Rewrite The Stars’. (The album’s lead single, ‘This Is Me’, winner of the 2017 Golden Globe Award for Best Original Song, was produced by Wells and mixed by Manny Marroquin, who had already been working on the song before Wells was hired for the film). The final Pro Tools mix session of ‘Rewrite The Stars’ contains 333 tracks — apparently a modest total by the standards of this album. Describing all these in detail would take up the rest of the magazine, but a broad overview from top to bottom (or left to right in the Mix window) begins with 23 grey tracks containing “all kinds of stuff, including stems from the original demo”, 30 tracks of orchestra (in blue), 10 tracks of demo mixes and other mix reference tracks, 43 tracks of drums and percussion, six tracks of bass and 63 tracks of guitar. These are followed by five tracks of piano, 16 tracks of strings, nine tracks of low percussion, five tracks of pads, another 13 tracks of percussion, 28 tracks of synth and piano, another five tracks of bass, six various, 10 more guitar tracks, a few miscellaneous tracks including a demo backing vocals track, eight tracks of Zac Efron lead vocals, nine tracks of Zendaya lead vocals, six tracks of Wells’ bussing, four auxiliary effect tracks, and four master tracks. Remarkably, nothing seems to be muted, so one wonders what choices Wells made. Surely he can’t have used everything?

“They shot all the movie footage to demos,” explains Wells, “so we needed to make sure the demos were in the session, with everything lined up sample accurate, to the nanosecond. So there are all the different stems that were used on set to play for the actors when the cameras were rolling. A lot of the stuff is from producer and composer Joseph Trapanese, including his routing, which is way over the top, with dozens of track outputs with names like ‘TRA’, ‘05RG’, ‘06LG’, ‘07PU’ and so on. Joe joked about it to me, and said that his routing also was designed to make it easier to create stems for the movie people. You solo one thing and it allows you to give them the stem they want in the super-particular way they need it. So instead of trying to replace it, I built around the routing that was already there, including his.

“Typically the sessions were given to me with audio arranged chronologically, with audio at the top from earlier sessions and at the bottom from later sessions. There always were multiple versions, and every time they gave me a new song it came with very detailed explanations. They shared all their knowledge and all their experiences, like this was the day we did it and this is what the weather was on that day, and this is why we did it, and so on. The reason is that they really turn every stone, and try everything. I had so much information coming at me, it was like I was hit by three or four tsunamis at once, almost all the time!

“I had to get used to the enormous sizes of the sessions and could not easily pare them down, because, as I mentioned, Justin was really forensic in his knowledge and was sitting next to me all the time. I thought I was into details, but he made me look like I’m blind. So I could not consolidate tracks, because he would always be chasing down the smallest nuances. The feeling of dealing with these enormous sessions became like conducting an orchestra. I think the fact that I played classical piano when I was a little boy was great training for an over-the-top project like this, because you learn to deal with a huge repertoire with long compositions and lots of information.”

The Pro Tools session is much too big to be usefully reproduced in these pages, and most of its 333 tracks are devoid of plug-ins. Major exceptions include the lead vocals by Zac Efron and Zendaya, and this screenshot shows some of the plug-ins Greg Wells used to try to generate a ‘pop’ vocal sound from a ‘soundtrack’ recording.

Tough Choices

In the case of ‘Rewrite The Stars’, Wells was told very early on that opinion within the production crew was split between two favourite versions, with no-one being able to work out how to arrive at an end result. The way Wells cut this particular Gordian knot would have made Alexander The Great proud. “The original version of this session was much bigger, with every different version of the song in it. I managed to whittle it down to some degree, including the material of their two favourite versions, which also were my favourite versions, with one being very orchestral and the other more pop-sounding with lots of programming and no orchestral stuff.

“Neither version was entirely right. So I started experimenting and experimenting to come to some sort of resolution, and one day when Justin and Benj were in New York, and I was on my own, I just thought: ‘Now that I have the lead vocals sitting where I want them, let’s see what it sounds like if I just play the two versions at the same time.’ To my happy surprise, they worked beautifully together. It actually sounded like they were part of the same track, which was a wonderful discovery, as these were two totally different versions done by different people in different years. I spent two more days adding some production elements, and mixed it from that point.”

Twin Lead

There are, in general, relatively few plug-ins on the individual tracks, with whole sections having no plug-ins at all. This is partly because many of the tracks are stems created by other producers, with effects already applied, and also because, says Wells, “There also aren’t many plug-ins on the orchestral stuff, because I like that to sound really open. For me the orchestra sound is about microphone choice and placement, and having great arrangements and great players in a great studio. That’s a sound I hardly ever mess with.”

The exceptions to this rule are to be found among the bottom 30 tracks of the session, where many plug-ins adorn the two lead vocals, Wells’ nine mix effect busses, titled ‘A’ to ‘F’ and ‘Verb/Delay’, ‘Drumcrush’ and ‘Fatso’, and his mix master bus. The structure of the lead vocal tracks by Zac Efron and Zendaya, and the plug-ins used on each of them, is almost identical. In each case it starts with a lead vocal audio track, which is then sent to a ‘Lead Vocals’ aux track and a ‘Lead Vocal Pultec’ parallel compression aux track. Both these aux tracks are sent to a ‘Lead Vocals Combiner’ aux, with the OEK Soothe plug-in in the insert and four sends to aux effects tracks with treatments titled ‘Crush’, ‘Verb’, ‘Delay’ and ‘Spread’. Unpacking this structure reveals that some of Wells’ venerated hardware had, after all, seen some action.

“The vocals were tuned, but I definitely got into tuning them myself. I’m very particular about tuning, and often request the raw vocals if I don’t like how they have been tuned. If needed I will pitch-shift or tune a tuned vocal again. I don’t like the sound of Melodyne, and instead use Waves Tune and Antares Auto-Tune. They’re both great, and each can do something the other can’t. I go back and forth between them all the time. Vocal timing is a big one too. I’m constantly moving things around timing-wise. Timing a phrase or a word a millisecond or two earlier or later can make a huge difference. I do this by hand, in the Pro Tools edit window. I’ll create a region and will move it by hand with the nudge button.

“The vocals also were recorded without much compression, and from the sound of it they probably had been singing five feet from the microphone — the stuff you do in movies, but you’d never do for a record. For me the sound of a modern record is the mic an inch in front of the singer, with compression. I also like the proximity effect. I wanted that modern sound on this song, as it’s so clearly a pop song. Zendaya is 21 years old, and definitely the pop factor in this movie. So I really was after a Serban Ghenea-type pop vocal sound, and it took a while to get there.

“To get the vocals to sound more energised, I used my interpretation of the Michael Brauer vocal thing, and sent each lead vocal to a range of different tube compressors, each of them on an analogue insert. These were my RCA BA6B, my mid-’60s UA 175b, my Retro Instruments 176 — it’s a new version of the 175b — the Chandler EMI RS124, the Unfairchild by Undertone Audio, the Retro Instruments Sta-Level and a Neve Shelford Channel. Every one of them is not compressing more than 2-3dB, and each of them sounds quite different. I blend the outputs of all of them to one vocal track, and I usually end up favouring one or two of the compressors over the others. This process winds up sort of EQ’ing the vocal in a really beautiful, musical way, without actually using EQ. I then print that back into the session, and this becomes the new vocal track, which are the tracks called ‘Zac Lead’ and ‘Zendaya Lead’.

“Each of these audio lead tracks goes into its own aux called ‘LeadVocals’, on which I have a signal chain on the inserts of the Waves Scheps 73 virtual Neve 1073, Crane Song Phoenix II tape emulation, and Waves Rvox. This is all based on Andrew Scheps’ technique, and sounds big and full and natural with a tiny bit of compression. Each audio lead track also goes to its own ‘LeadVocalPultec’ aux track, which has a signal chain of a UAD Pultec EQP-1A, a Waves CLA-2A, a second Pultec EQP-1A and again the Scheps 73, and here the EQ has more of a disco smile, with lots of 100Hz and top end added, and tons of compression.

“I blend these two tracks until they sound right, and then they go to the ‘Lead Vocals Combiner’ aux, which has a new plug-in called Soothe, made by a Finnish company called OEK Sound. It’s really amazing. When you use it sparingly, and set it to the oversample option, the effect is phenomenal. It just takes out all the ugly harmonic stuff. Microphones are weird and unnatural, and some singers’ voices respond beautifully to a microphone, whereas others don’t. It’s like some people are photogenic and some people are not. We have all read about people like Mutt Lange getting his engineer to EQ every single syllable of a vocal, and I have done that as well. You can make a vocal sound more natural like that, as if there’s no microphone and the singer is right there in front of you. The Soothe does the same thing, but it can make 50 different moves in one syllable. It is brilliant. It is really incredible. It’s on both lead vocals, and it really helped to give me the pop vocal sound that I wanted on both of the singers.

“Each ‘Lead Vocal Combiner’ track has four sends, which go to a ‘Vocal Crush’ aux track, with a UAD 1176 really slammed and a fast release, a ‘Vocal Verb’ track with the McDSP Filterbank EQ and some nice, open reverb from a UAD Pure plate, a ‘Vocal Delay’ track with the Waves H-Delay, and a ‘Spread Vox’ aux with the Eventide H910 Harmonizer for some spread. I confidently use the same signal chain structure and plug-ins on both vocals, because my setup is a catch-all. When I did my VoiceCentric plug-in for Waves, they asked how a male vocal chain can also be used for a female vocal chain and I explained that it can: it is just a framing of the vocals in the track. I actually use VoiceCentric all the time, but the reason that it’s not in this session is again because the movie world wanted me to stem out dry vocal stems and stems with each individual effect, and they lose their minds and yell if they don’t get it! So I had to have a tweak-y, lifting-the-hood manoeuvrability with the vocal effects that the VoiceCentric intentionally doesn’t have, because I wanted that plug-in to sound incredible, but also be really simple to use with a minimum of control options.”

On The Busses

Underneath the vocal tracks are Wells’ nine mix effect aux tracks, or busses. Wells: “This is inspired by Michael Brauer’s multiple bus approach. I think he has four or five, but I wound up with nine mix busses in this session. All the lead vocals go to ‘Bus A’, which has a very small amount of UAD Neve 33609 compression and a little bit of UAD Pultec EQP-1A, adding 2dB at 16kHz and a little bit of 100Hz. All drums and bass go to ‘Bus B’, which has an Empirical Labs Arouser, followed by Fabrice Gabriel’s amazing Eiosis AirEQ plug-in, which I adore. I use it to tweak some key frequencies for some more presence.

“‘Bus B’ also has some sends to three other busses. The first send goes to the Fatso bus, which has UAD Fatso Jr compression. It’s a glorious-sounding plug-in, which gives more weight in the lower mids and the low end. The second send goes to ‘Bus E’, which has the UAD Thermionic Culture Vulture and the UAD LA-2A. I have an outboard Culture Vulture, but the UAD version sounds fantastic. The LA2A does some slow optical compression to give it a little bit of that Alan Moulder bottom end, which is an amazing way to give your drums and bass extra depth. The third send goes to the Drum Crush bus, which has the UAD Fairchild 670 compressor, which slams the signal quite hard, with a release that is set to the second fastest setting. I just blend that in to taste.

“‘Bus C’ has the really incredible Stephen Slate FG-MU, which is their version of the Fairchild, the Slate Virtual Mix Rack, and then the UAD SPL Vitalizer. Again, I also have the hardware version, but the plug-in sounds great. I send guitars to ‘Bus C’, sometimes piano, anything that’s pretty natural-sounding but needs a little bit more zing. I don’t use ‘Bus D’ very much. It has an old Focusrite compressor plug-in and the Avid AIR Stereo Width. That is another Michael Brauer trick. He always has a bus with a widener. I know there are many other stereo wideners, but I think that the AIR one is great.

“I also send orchestral drums to ‘Bus E’, which I mentioned above, which I use for anything that needs that glorious Alan Moulder bottom end. Strings and anything orchestral goes to ‘Bus F’, which has a little bit of UAD SPL Vitalizer, with the tiniest bit of stereo expansion, a tiny bit of the Process knob, and adding some upper mids and a little bit of low end. After ‘Bus F’, the next bus is the ‘Verb/Delay’ return. Any effects that I don’t want any [additional] processing on will show up there — the four vocal effects aux tracks go through this, for example. Then there’s the ‘Drum Crush’ bus, with the UAD Fairchild 670, and finally the ‘Fatso’ bus, with the Fatso Jr and the LA-2AG. There’s a feed track that controls the level going into that ‘Fatso’ bus, which is an Andrew Scheps trick. I got that directly from him and I love it.”

Greg Wells is also not shy of using plug-ins on his master bus!

Master Processing

Wells: “All these nine busses are sent to the ‘Mix Bus Level’ track, which feeds the ‘Mix Bus’ track with all the plug-ins. This is another thing I got from Andrew. The ‘Mix Bus’ signal chain starts with the UAD Neve 33609, and we both love the sound of that on the mix bus. It can provide some really magical-sounding glue, but it also is very grabby, and if you send too much level to it, it just sounds ridiculous. You have to pull the input signal down until it is almost doing nothing, so that is what I am doing with the ‘Mix Bus Level’ track. After the 33609 I have the Black Box Analogue Design HG-2 tube compressor, which sounds great. I added a little bit of all the different choices that it offers, and I backed off the output level. I do the latter as a rule, because something that sounds louder instantly sounds better, so I set it so that when I bypass a plug-in the level is the same. This means I can make a genuine comparison, and make sure it actually sounds better at the same level.

“Next is the Brainworx bx_digital V3, which is a wonderful surgical EQ, which also has a great stereo width knob, and then I again have the Slate Virtual Channel on a Neve setting, which gives some more definition down in the kick-drum area. I have the Slate Virtual Tube Collection, and I have the New York Tube selected with a little bit of saturation and the mix knob set to 80 percent. After that we have the UAD Brainworx bx_subsynth, which is adding some super, ultra-low movie rumble, with the mix knob set to 26 percent. After that I have a plug-in that I don’t use that often, the UAD Precision Maximizer. It is a beautiful plug-in, but I don’t like the sound of the limiter so I always turn that off. For the rest I have it set to three bands, and the mix is at 100 percent and the shape is set to 63 percent.

“I had not installed [iZotope] Ozone 8 yet, so I am still using 7 in this session, and it does some very gentle stuff here. The Maximiser is bypassed, and there’s a little bit of a multiband dynamic thing happening, with the parallel blend set to 52 percent, so it’s not doing very much. There also is some low roll-off in the EQ at 40Hz and some upper mid boost. Finally, there’s the Sonnox Oxford Limiter. I love the Enhance fader on that, and it is set to 59 percent. I also do a slightly eccentric thing, which is to blend in parallel compression on the entire mix bus. I have a send from the Mix Bus to a ‘mix parallel’ bus next to it, on which I have one of these Focusrite compressor plug-ins, and I’m squashing that pretty hard. I blend that in with the main Mix Bus, with the ‘mix parallel’ set to -17dB.

“There are well-known mixers who say that you should not put a lot of stuff on your mix bus, but I put tons on, and some others do as well. Everything I put on does a tiny little bit. It is cumulative. I read a quote by Daniel Lanois many years ago in which he said that he thinks an EQ sounds better if you send a lot of stuff through it. Rather than EQ’ing individual elements, he likes to send 15 different things to the same EQ. For him that helped get the sound that he was hearing in his head. I know exactly what he means. I think I am very influenced by the fact that my very first studio was a four-track Fostex cassette machine. I still have it at my studio to remind me how this all started. Because I had only four tracks I always had to process the end result, and I am still influenced by that. Everyone has their own way of arriving at the finish line, and for me not being conservative with processing on the mix bus is key.”

The Greatest Show On Earth

Having arrived at the finish line at the end of 2017, Wells recalls that he had no idea how the movie and the soundtrack album would be doing. With The Greatest Showman being director Michael Gracey’s first feature movie, and the whole project having taken seven years because the studios considered taking on a musical high risk, the movie and soundtrack album could have disappeared without trace. However, the film’s success pushed the soundtrack album to the top spot in the UK and the US, and it also was the iTunes number one in a whopping 65 countries.